[unreadable] Each year, a new influenza strain spreads across the nation, through cities interconnected by air and ground transportation. Consequently, there is potential for disastrous consequences, should an influenza pandemic strain emerge. In 1918, the "Spanish flu" pandemic claimed over 675,000 lives in the United States, and tens of millions worldwide. Because the last was in 1969, there is mounting concern over a recurrence; and the alarming spread of the H5N1 highly pathogenic avian influenza virus among poultry in Southeast Asia only heightens this concern. Surprisingly, there is little empirical data on how influenza spreads through cities, regions, nations and across the globe. Targeted prevention strategies could, however, be informed by enhanced understanding of local and large scale patterns. We will leverage an NLM-funded automated advanced disease surveillance system (AEGIS) to empirically measure the key determinants of influenza spread. Our first aim is to develop models of national influenza spread to inform public health preparedness strategies for epidemic and pandemic flu. We will model the national spread of the yearly influenza epidemic using network driven methods so as to understand the impact of multivariate factors, including population movement and environmental conditions. We will map the yearly pattern of spread and identify urban hubs that represent major network pathways. We expect to identify sentinel cities that should be the focus of control strategies which could include targeted vaccination, travel advisories, and flight bans. Our second aim is to develop spatial models of local influenza spread in metropolitan areas to identify targets for surveillance and control. We will develop empirically based spatial models of local influenza spread across major metropolitan areas and identify recurring hotspots of risk using surveillance data. We will exploit differences in spatial patterns of infection in major metropolitan areas to evaluate geographic and demographic risk factors that drive infection in different parts of the country. For both aims, our models and methods will be implemented in the AEGIS surveillance system for real-time monitoring of influenza activity. We will develop methods for effective linkage of influenza surveillance to prevention and control strategies. These methods have broad application for disaster preparedness for events caused by both naturally-occurring and deliberate introductions of infectious agents. [unreadable] [unreadable] [unreadable] [unreadable]